Rev 2 | Details | Compare with Previous | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 2 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_q15_to_q7.c |
||
| 4 | * Description: Converts the elements of the Q15 vector to Q7 vector |
||
| 5 | * |
||
| 6 | * $Date: 27. January 2017 |
||
| 7 | * $Revision: V.1.5.1 |
||
| 8 | * |
||
| 9 | * Target Processor: Cortex-M cores |
||
| 10 | * -------------------------------------------------------------------- */ |
||
| 11 | /* |
||
| 12 | * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved. |
||
| 13 | * |
||
| 14 | * SPDX-License-Identifier: Apache-2.0 |
||
| 15 | * |
||
| 16 | * Licensed under the Apache License, Version 2.0 (the License); you may |
||
| 17 | * not use this file except in compliance with the License. |
||
| 18 | * You may obtain a copy of the License at |
||
| 19 | * |
||
| 20 | * www.apache.org/licenses/LICENSE-2.0 |
||
| 21 | * |
||
| 22 | * Unless required by applicable law or agreed to in writing, software |
||
| 23 | * distributed under the License is distributed on an AS IS BASIS, WITHOUT |
||
| 24 | * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
||
| 25 | * See the License for the specific language governing permissions and |
||
| 26 | * limitations under the License. |
||
| 27 | */ |
||
| 28 | |||
| 29 | #include "arm_math.h" |
||
| 30 | |||
| 31 | /** |
||
| 32 | * @ingroup groupSupport |
||
| 33 | */ |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @addtogroup q15_to_x |
||
| 37 | * @{ |
||
| 38 | */ |
||
| 39 | |||
| 40 | |||
| 41 | /** |
||
| 42 | * @brief Converts the elements of the Q15 vector to Q7 vector. |
||
| 43 | * @param[in] *pSrc points to the Q15 input vector |
||
| 44 | * @param[out] *pDst points to the Q7 output vector |
||
| 45 | * @param[in] blockSize length of the input vector |
||
| 46 | * @return none. |
||
| 47 | * |
||
| 48 | * \par Description: |
||
| 49 | * |
||
| 50 | * The equation used for the conversion process is: |
||
| 51 | * |
||
| 52 | * <pre> |
||
| 53 | * pDst[n] = (q7_t) pSrc[n] >> 8; 0 <= n < blockSize. |
||
| 54 | * </pre> |
||
| 55 | * |
||
| 56 | */ |
||
| 57 | |||
| 58 | |||
| 59 | void arm_q15_to_q7( |
||
| 60 | q15_t * pSrc, |
||
| 61 | q7_t * pDst, |
||
| 62 | uint32_t blockSize) |
||
| 63 | { |
||
| 64 | q15_t *pIn = pSrc; /* Src pointer */ |
||
| 65 | uint32_t blkCnt; /* loop counter */ |
||
| 66 | |||
| 67 | #if defined (ARM_MATH_DSP) |
||
| 68 | |||
| 69 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 70 | q31_t in1, in2; |
||
| 71 | q31_t out1, out2; |
||
| 72 | |||
| 73 | /*loop Unrolling */ |
||
| 74 | blkCnt = blockSize >> 2U; |
||
| 75 | |||
| 76 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 77 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 78 | while (blkCnt > 0U) |
||
| 79 | { |
||
| 80 | /* C = (q7_t) A >> 8 */ |
||
| 81 | /* convert from q15 to q7 and then store the results in the destination buffer */ |
||
| 82 | in1 = *__SIMD32(pIn)++; |
||
| 83 | in2 = *__SIMD32(pIn)++; |
||
| 84 | |||
| 85 | #ifndef ARM_MATH_BIG_ENDIAN |
||
| 86 | |||
| 87 | out1 = __PKHTB(in2, in1, 16); |
||
| 88 | out2 = __PKHBT(in2, in1, 16); |
||
| 89 | |||
| 90 | #else |
||
| 91 | |||
| 92 | out1 = __PKHTB(in1, in2, 16); |
||
| 93 | out2 = __PKHBT(in1, in2, 16); |
||
| 94 | |||
| 95 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
||
| 96 | |||
| 97 | /* rotate packed value by 24 */ |
||
| 98 | out2 = ((uint32_t) out2 << 8) | ((uint32_t) out2 >> 24); |
||
| 99 | |||
| 100 | /* anding with 0xff00ff00 to get two 8 bit values */ |
||
| 101 | out1 = out1 & 0xFF00FF00; |
||
| 102 | /* anding with 0x00ff00ff to get two 8 bit values */ |
||
| 103 | out2 = out2 & 0x00FF00FF; |
||
| 104 | |||
| 105 | /* oring two values(contains two 8 bit values) to get four packed 8 bit values */ |
||
| 106 | out1 = out1 | out2; |
||
| 107 | |||
| 108 | /* store 4 samples at a time to destiantion buffer */ |
||
| 109 | *__SIMD32(pDst)++ = out1; |
||
| 110 | |||
| 111 | /* Decrement the loop counter */ |
||
| 112 | blkCnt--; |
||
| 113 | } |
||
| 114 | |||
| 115 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here. |
||
| 116 | ** No loop unrolling is used. */ |
||
| 117 | blkCnt = blockSize % 0x4U; |
||
| 118 | |||
| 119 | #else |
||
| 120 | |||
| 121 | /* Run the below code for Cortex-M0 */ |
||
| 122 | |||
| 123 | /* Loop over blockSize number of values */ |
||
| 124 | blkCnt = blockSize; |
||
| 125 | |||
| 126 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 127 | |||
| 128 | while (blkCnt > 0U) |
||
| 129 | { |
||
| 130 | /* C = (q7_t) A >> 8 */ |
||
| 131 | /* convert from q15 to q7 and then store the results in the destination buffer */ |
||
| 132 | *pDst++ = (q7_t) (*pIn++ >> 8); |
||
| 133 | |||
| 134 | /* Decrement the loop counter */ |
||
| 135 | blkCnt--; |
||
| 136 | } |
||
| 137 | |||
| 138 | } |
||
| 139 | |||
| 140 | /** |
||
| 141 | * @} end of q15_to_x group |
||
| 142 | */ |